Metatranscriptome analysis pipeline

[Francy87]

Hello there!
I've just started analysing bacterial metatranscriptome Illumina data. I'm a bit confused on the pipeline to follow. That's what I did:
-I used FastQC to have a general overview of my sequences quality, then I removed the adaptors using the FastX toolkit;
-I assembled the data using Velvet and then Oases, obtaining a file with transcripts.
Now I'm not sure about how going on... I thought to use BLAST for annotation and then mapping the files from each sample individually to the assembled transcriptome using BWA... But then I don't know how proceed... I'd like to have something like this:
GeneID SampleID1 SampleID2 .......

gene1 %abundance %abund
gene2 %abund %abund
....
Then use this table to build some of those awesome plots like heatmaps...
Is there a software that produce an output like this?? Then, does anyone have comments/suggestions on the workflow I used?
Thanks a lot
Francesca

[Blahah404]

You should read the tutorials in the khmer documentation, and then make use of the khmer scripts.

Briefly, you do your read trimming and adapter removal, then you perform digital normalisation to reduce the redundancy in your read set, then you partition the de-brujn graph to separate out different organisms, and then you do the assembly.